INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SKELETAL MUSCLE CAPILLARIES form an elaborate network, the abundance and structure of which, in part, help regulate muscle perfusion and O₂ delivery/exchange. Any structural or functional impairments at this level will impact blood-myocyte O₂ and substrate exchange. Although a decreased or maldistributed cardiac output, reduced skeletal muscle blood flow, and intrinsic abnormalities in skeletal muscle have been implicated in chronic heart failure (CHF)-induced skeletal muscle dysfunction at rest and during exercise (5, 6, 17, 18, 20, 27, 34, 38, 42, 44), skeletal muscle capillary hemodynamics have not been quantified. Gaining an understanding of microcirculatory structure as well as red blood cell (RBC) distribution within this network will lend insight into the muscle O₂ diffusing capacity (Dm_O2) and thus O₂ exchange (7) in this pathophysiological state. Specifically, the theoretical calculations of Federspiel and Popel (7) suggest that the number of RBCs adjacent to the muscle fiber at any given instant is a critical index of O₂ exchange. Thus, as the ratio of RBC surface area to myocyte surface area increases, Dm_O2 is expected to increase proportionally. Capillary "tube" hematocrit is the primary determinant of RBC surface area, and whether this is altered in CHF has not been determined.

CHF is associated with a decreased skeletal muscle perfusion at rest (6, 17, 27, 38, 42, 44) as well as an increased O₂ extraction (37). In addition, a reduced skeletal muscle mass (18, 20, 37) and an increased percentage of type IIb fibers, a decreased proportion of type I fibers, and a preferential atrophy of type IIb fibers (5, 18, 20, 37) are manifest in this disease state. With respect to microvascular structure, skeletal muscle from CHF patients may exhibit a slight decrease in capillary density (20) or capillary length density (5) or no change in either capillary density (18) or capillary-to-fiber ratio (37). In one recent investigation in the rat model of CHF, plantaris muscle capillary density measured via morphometric techniques was unchanged 7 wk after myocardial infarction, reflecting a modest fiber atrophy in combination with capillary involution (43). This latter investigation also reported no change in capillary orientation and diameter (43). We consider it likely that, 7 wk after infarct (as studied in this investigation), the reduced blood flow to the periphery in CHF at rest (6, 17, 27, 38, 42, 44) is associated with an altered profile of muscle capillary hemodynamics that occurs in the absence of marked capillary structural alterations. Designation of how the RBC hemodynamics are altered within the microcirculation will lend mechanistic insight into alterations of Dm_O2 and O₂ exchange in CHF. If O₂ extraction is determined by the ratio of Dm_O2 to blood flow (), as suggested by Roca et al. (31), where O₂ = O₂ · [1  exp(Dm_O2/)] and percent O₂ extraction = O₂/O₂ = 1  exp(Dm_O2/) (where O₂ is O₂ consumption, O₂ is O₂ delivery, and  is the slope of the O₂ dissociation curve in the physiologically relevant range), then it is likely that the CHF-induced increases in O₂ extraction result from a reduced RBC velocity and, thus, increased RBC transit time within individual capillaries consequent to reduced bulk blood flow. However, for any given reduction in muscle blood flow in CHF, the resultant change in RBC velocity and capillary residence time will depend on the proportion of capillaries supporting RBC flux and the hemodynamics within those capillaries (i.e., RBC flux and velocity as well as tube hematocrit). As mentioned above, whether altered capillary tube hematocrit changes Dm_O2 under these circumstances is not known.

Thus the purpose of the present investigation was to determine 1) the effect of CHF on skeletal muscle capillary structure (i.e., capillary diameter and lineal density) and orientation (i.e., proportion of total capillary length due to tortuosity and branching) measured in vivo and 2) skeletal muscle microcirculatory hemodynamics (i.e., proportion of RBC-perfused capillaries, RBC flux and velocity, and capillary tube hematocrit) in CHF. Specifically, we tested the hypothesis that at resting sarcomere length (i.e., ~2.7 µm) the impaired muscle blood flow in CHF will 1) reduce the proportion of capillaries supporting RBC flow while maintaining RBC velocity and flux or 2) decrease RBC velocity and flux while maintaining a similar absolute number of RBC-perfused capillaries.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental animals. Female Wistar rats (initially ~225 g) were used in this investigation. All procedures were approved by Kansas State University animal-handling guidelines. Rats were divided randomly into sham and CHF groups, in which CHF was induced by surgically ligating the left main coronary artery and thereby producing a moderate-to-large myocardial infarction, as described previously (26). Briefly, all rats were anesthetized with 3% halothane, intubated, and placed on a respirator (model 680, Harvard). Anesthesia was maintained using a 2% halothane-O₂ mixture. A left thoracotomy was performed in the fifth intercostal space, the pericardium was opened, and the heart was exposed. In CHF rats the left main coronary artery was ligated (6-0 suture) ~1 mm distal to the edge of the left atrium. In sham rats the coronary artery was not ligated. After surgery, all rats were housed individually in 6 × 9 in. cages to restrict activity (35).

Muscle preparation. The spinotrapezius muscle lies in the middorsal region of the rat; it originates in the lower thoracic and upper lumbar region and inserts on the spine of the scapula. The spinotrapezius muscle was prepared surgically using methods previously described (8, 29). Surgery was performed with minimal fascial disturbance to minimize tissue damage (i.e., all vascular and nervous connections remained intact) and any associated microcirculatory consequences (8, 23). Briefly, the rat was placed on a circulation-heated (38°C) Lucite platform, and the spinotrapezius muscle was superfused continuously (Microcirculator U3-7A, Julabo, Schwarzwald, Germany) with a Krebs-Henseleit bicarbonate-buffered solution equilibrated with 95% N₂-5% CO₂ (39). The exposed dorsal surface of the spinotrapezius muscle was protected with plastic wrap (Dow, Indianapolis, IN) while the muscle was sutured (6.0 silk) at five equidistant positions around the caudal periphery to a thin wire horseshoe manifold (39). Thereafter, all exposed surfaces were covered with plastic wrap or bathed in the physiological solution. The manifold was attached with a swivel to a muscle-stretching apparatus that permitted precise, systematic, and uniform length changes of the entire spinotrapezius muscle along its principal fiber longitudinal axis.

Intravital video microscopy. Images were obtained using an intravital video microscope (model OVM-1000NM, Olympus, Tokyo, Japan) equipped with a noncontact, illuminated lens and viewed on a high-resolution (>600 lines) color monitor (Trinitron PVM-1954Q, Sony, Ichinoniya, Japan). Final screen magnification was ×1,500, as confirmed from initial calibration of the system by means of a stage micrometer (model MA285, Meiji Techno). Images were time referenced by frame, and fields were stored via videocassette recorder (model BR-S822U, JVC, Elmwood Park, NJ) on Super VHS cassettes (S-VHS MASTER XG, JVC) for subsequent off-line analysis. The microvascular field (270 × 210 µm) as viewed on the monitor typically contained four to six muscle fibers for the sham and six to eight for the CHF animals. The muscle was transilluminated using a fiber-optic light source (Fiber-Lite, Dolan-Jenner Industries, Lawrence, MA) at an incident light angle that provided clear visualization of the A-bands within one- to two-thirds of the muscle fibers. Thus simultaneous measurements of sarcomere length, capillary geometry, and flow dynamics could be obtained without compromising the hemodynamic behavior of the preparation.

Experimental design. The spontaneously shortened spinotrapezius muscle assumes a sarcomere length of ~2.2 µm (29). With use of the manifold attached to the muscle, sarcomere length was increased to ~2.7 µm, as verified by direct on-screen measurements (see Capillary and fiber structural data). Typically, up to three microvascular fields demonstrating good clarity and located approximately medial to the arteriolar and venular ends of the capillaries were observed for 90-120 s each at this sarcomere length for further off-line analysis. Differences in microvessel and fiber clarity between muscle regions in this preparation appeared to result from the presence of the overlying fascial sheath rather than the characteristics of the underlying tissue or its capillary organization. Despite this, we chose not to remove this connective tissue because of the probability of causing substantial tissue damage (23). The structures of primary interest were the capillaries with their associated RBCs and muscle fiber sarcomeres (A-bands). If these were not clearly visible within a given region, the sample was not selected for subsequent analysis. We utilized the following working definition for capillaries: those vessels intermediate between the terminal arteriole and the collecting venule, which were <10 µm in diameter. The structural arrangement of the capillary bed in the spinotrapezius (36) enables the positioning of the optical window (270 × 210 µm) so that it views the central portion of the capillary bed in a region where terminal arterioles or collecting venules are very rare. On those few occasions where the identity of a given vessel was ambiguous, the vessel was tracked anterograde or retrograde to establish its identity. Although the presence of single-file RBCs was not a criterion for identification of a capillary, all vessels identified as such did exhibit single-file RBCs. At the end of each experiment, the field was shifted to enable visualization of at least one arteriole, and adenosine (10⁴ M) was applied topically to the muscle. Adenosine induced a profound arteriolar vasoactive response in control animals. Compared with sham rats, this compound appeared to have a less pronounced effect in all but two CHF animals. Experimental duration was 1-1.5 h, during which up to 1.5 ml of sterile isotonic saline were infused intra-arterially in 0.1-ml bolus increments to counteract dehydration.

Capillary and fiber structural data. Each viewing field was screened for confirmation of sarcomere length and clarity. Muscle sarcomere length was determined from sets of 10 consecutive in-register sarcomeres (i.e., distance between 11 consecutive A-bands) measured parallel to the muscle fiber longitudinal axis. Within the resolution of the monitor, this method of measuring sarcomere length by averaging 10 consecutive sarcomeres allows for sarcomere length to be set at 0.05-µm increments. This procedure was performed, in duplicate, on each muscle fiber and on every muscle fiber within the viewing field where sarcomeres were visible to obtain a mean sarcomere length for each viewing field. If interfiber sarcomere length variance was >0.1 µm, the field was discarded. Consequently, within each preparation, only that field within 2.7 ± 0.1 µm sarcomere length, which portrayed the best overall clarity, was chosen for further study. These fields were traced directly from the video monitor screen onto acetate paper. The details traced included the muscle fiber boundaries, ~25 A-bands within each fiber, and the lower margin of the capillary endothelium, where it was visible continuously. For each muscle fiber in which both sarcolemmal boundaries were visible on-screen, the apparent fiber width perpendicular to the longitudinal muscle fiber axis was measured, and associated capillaries were counted. These values were used to calculate lineal density (i.e., the number of capillaries per unit muscle width). To ensure that the count of capillaries per fiber was not overestimated by inclusion of vessels at the periphery of the screen (edge effect), capillaries were counted between the midpoints of adjacent muscle fibers. This is akin to the "forbidden line" rule in morphometry. Where the capillary endothelium was clearly visible on both sides of the lumen, capillary luminal diameter was measured with hand-held calipers at two random locations, in duplicate, per capillary, and the mean value was recorded. In the hands of an experienced technician, this procedure was accurate to ±0.25 mm (i.e., ±0.17 µm at ×1,500 magnification).

Capillary geometry. Capillary geometry was determined using a variation of the technique described by Batra and Rakusan (2), providing a value analogous to the capillary anisotropy coefficient [i.e., c(K,0)] obtained by ex vivo morphometric methods (22). From the acetate paper tracing, each capillary was first labeled for organizational purposes, and the beginning and end points of each capillary were denoted. These points included as much of the capillary length as possible with the prerequisite that the capillary remain in focus for the entire length measured. A thin, flexible wire was used to trace the actual path of each capillary along the lower visible edge of the endothelium from beginning to end points. In those capillaries where capillary interconnections were encountered, one-half of that capillary interconnection length was added to the measured capillary length. These capillary interconnections (usually 30-60 µm long) are found generally between two adjacent capillaries. The straightened wire length and one-half of any associated interconnection length allowed for measurement of the length of a given capillary. This provided a measurement of absolute capillary length within each screen. Next, the corresponding length (of each capillary) along the muscle fiber longitudinal axis was obtained by measuring the straight-line distance strictly parallel to the muscle fiber longitudinal axis between the previously established visible beginning and end points of each capillary. To obtain the proportion of the additional capillary length resulting from capillary tortuosity, branching, and deviations of alignment with the fiber axis [analogous to c(K,0)], the actual capillary path length and length of the associated capillary interconnections were divided by the straight-line length along the muscle fiber longitudinal axis.

Capillary hemodynamic data collection. RBC flow was observed in real time and also with use of video frame-by-frame playback techniques. Each capillary visible in a given field was observed for a random, continuous 60-s period and placed into one of three categories: 1) continuous RBC flow, 2) intermittent RBC flow, or 3) impeded RBC flow (i.e., stationary or no visible RBCs). This criterion was the basis for the quantification of the percentage of capillaries supporting continuous RBC flow, i.e., category 1. In addition, intracapillary RBC velocity (V_RBC) was determined in all capillaries in which RBCs could be followed over several frames. In those few capillaries in which RBC flow was arrested but resumed within the (60-s) sampling period, RBC velocity was determined when flow resumed. For these capillaries the mean RBC velocity (and RBC flux, F_RBC) was determined taking into account the percentage of the 60-s window in which RBC flow was stopped. RBC flux was also measured by counting and recording the number of RBCs that passed an arbitrary point on each capillary as viewed frame-by-frame over a 15-s period. On a few occasions the RBCs could not be counted over the entire 15-s period. In these instances the usual period was subsampled and mean RBC flux was averaged over 15 s.

Capillary tube hematocrit assessment. In each muscle, capillary tube hematocrit (Hct_t) was measured as follows where RBC volume was taken to be 61 µm³ (1), capillaries were assumed to be circular in cross section, F_RBC is RBC flux, and d_c is capillary luminal diameter. One average hematocrit value was calculated for each muscle.

Infarct size determination. The heart of each animal was removed at the completion of each experiment and immersed in Formalin for >48 h for fixation. Subsequently, the LV of each heart was cut into four transverse sections in parallel with the atrioventricular groove. The four sections were then dehydrated in alcohol, cleaned with xylene, and set in paraffin. Transverse sections (7 µm thick) were cut, mounted, and stained with Masson's trichrome stain (hemotoxylin was omitted to provide maximal discrimination between fibrous areas of infarct and viable muscle). The sections were magnified and projected, and the size of the infarcted areas was determined by planimetry, as described previously (28).

Statistical analysis. Values are means ± SE. All data were checked for normality of distribution by means of the Kolmogorov-Smirnov technique with use of SigmaStat. Data were regressed using a standard least-squares regression technique. Sham and CHF groups were compared using a Student's t-test. P  0.05 was accepted.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Successful experiments were completed on six of eight CHF rats and seven of nine sham rats. The infarct sizes in the CHF rats were 32.7 ± 4.8% (range 12.7-45.1%) of the LV endocardial circumference. Body mass of sham rats was not statistically different from that of CHF rats (269 ± 4 and 266 ± 5 g for sham and CHF, respectively, P > 0.05). Anatomic data and blood pressure measurements for sham and CHF rats are shown in Tables 1 and 2. There were no differences between groups for LV mass-to-body mass ratio, mean arterial pressure, or LV peak systolic pressure. However, LV end-diastolic pressure and right ventricular mass-to-body mass ratios were elevated in CHF animals, whereas LV developed pressure was reduced. In addition, lung mass-to-body mass ratios tended to be greater (P = 0.12) in CHF than in sham rats (Table 1).

Muscle and capillary structural changes. Spinotrapezius muscle length was set to approximately the same resting sarcomere lengths for both groups (2.7 ± 0.1 and 2.6 ± 0.1 µm for sham and CHF, respectively). Despite no difference in body mass, the spinotrapezius muscle fiber width was reduced in CHF rats (51.3 ± 1.9 and 42.6 ± 1.4 µm for sham and CHF, respectively, P < 0.01). With the assumption that the fibers are circular (or at least have no preferential cross-sectional orientation), this corresponds to a reduction in mean fiber cross-sectional area of ~30%. Capillaries of CHF rats were slightly straighter than those of sham rats; i.e., the proportion of capillary length due to tortuosity and branching was less in CHF than in sham rats (7 ± 1 vs. 10 ± 1%, P < 0.01). The mean capillary diameter in the CHF rats was not different from that in sham rats (5.4 ± 0.1 and 5.2 ± 0.1 µm for sham and CHF, respectively, P > 0.05). The reduced muscle fiber width found in CHF rats resulted in an increased lineal density of capillaries compared with sham rats (49.7 ± 5.9 and 35.3 ± 1.9 capillaries/mm for CHF and sham, respectively, P < 0.05).

Hemodynamics. Perfusion pressures in CHF and sham rats were not different during the actual intravital data acquisition (98 ± 9 and 96 ± 6 mmHg for sham and CHF, respectively, P > 0.05). The percentage of capillaries supporting continuous RBC flow was less in CHF rats, whereas the percentage of capillaries with intermittent RBC flow was greater (Table 3). There was no difference between sham and CHF rats with respect to the percentage of capillaries in the "impeded" category. The most pronounced reduction in the percentage of continuously RBC-perfused capillaries occurred in those rats with the largest infarcts. Specifically, in the three CHF animals with infarcts <35%, the proportion of capillaries maintaining continuous RBC flow was 78 ± 1% vs. 56 ± 2% in the three animals with infarcts >35% (P < 0.001). The decreased proportion of continuously flowing RBC-perfused vessels in CHF rats resulted in a similar lineal density for continuously flowing RBC-perfused capillaries between groups (30.5 ± 1.5 and 31.8 ± 5.7 capillaries/mm for sham and CHF, respectively, P > 0.05). The lineal density of capillaries supporting continuous RBC flow was negatively correlated with infarct size in CHF rats (Fig. 1).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Relationship between lineal density of red blood cell (RBC)-perfused capillaries with continuous flow in spinotrapezius muscle and infarct size in rats with chronic heart failure (CHF).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Relative frequency histogram of RBC flux and RBC velocity in skeletal muscle capillaries supporting flow for sham (S) and CHF rats. At left, arrows represent mean RBC flux, which was reduced in CHF skeletal muscle (P < 0.01). At right, arrows represent mean RBC velocity, which was also reduced in CHF compared with sham rats (P < 0.001).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Proportion of capillaries supporting continuous RBC flow vs. RBC velocity in spinotrapezius muscle of sham and CHF rats. There was a significant correlation (y = 5.6 + 0.42x, r = 0.81) between these 2 variables within CHF group only (). Three  symbols at the top of the regression line (near control values) represent mean values in 3 rats with infarcts <35%;  at bottom represent 3 CHF rats with infarcts >35%.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Relationship between mean RBC velocity and flux in capillaries supporting continuous RBC flow for sham and CHF animals.

1

1

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Previous measurements of reduced muscle bulk blood flow in CHF (6, 17, 27, 38, 42) provide no information regarding the distribution of that flow within the capillary network. Specifically, it was not known whether CHF reduces flow uniformly in all RBC-perfused capillaries or whether the proportion of RBC-perfused capillaries is altered. This information is critical if we are to understand the effect of CHF on muscle capillary RBC distribution and O₂ exchange. This investigation has demonstrated, for the first time, the effect of CHF on skeletal muscle capillary hemodynamics. Specifically, 7 wk after ligation of the left main coronary artery, the proportion of capillaries exhibiting zero RBC velocity and flux at some time within the sampling window increased from 13 to 33%. Furthermore, in those capillaries supporting continuous RBC perfusion, CHF markedly reduced RBC velocity (43%) and flux (56%) in the face of an unchanged capillary tube hematocrit. Despite a small (although significant) reduction in c(K,0), the substantial decrease in capillary RBC velocity elevated the estimated capillary RBC residence time, thereby enhancing the potential for O₂ off-loading in CHF muscles. That capillary tube hematocrit was unaltered suggests that any change in Dm_O2 consequent to the CHF condition occurs at a site distal to the RBC.

Sham vs. control. Within the pertinent literature, the tacit assumption is that the sham-operated rat provides the most appropriate control for its CHF counterpart (24-27). With respect to the key structural (i.e., body mass, spinotrapezius fiber diameter, capillary diameter, and geometry) and hemodynamic variables, no differences were evident between published control values (3, 11, 13, 14, 29, 33) and values reported here for "sham" controls.

Quantification of CHF. The CHF condition in the present investigation was induced using an established model of myocardial infarction in the rat (6, 24-28, 34). Compared with sham, CHF rats exhibited elevated right ventricular mass-to-body mass ratio and LV end-diastolic pressure values as well as a reduced LV developed pressure (Tables 1 and 2). Also, there were trends toward decreased mean arterial pressure and LV systolic pressure and increased lung mass-to-body mass ratio in the CHF rats. These anatomic and hemodynamic characteristics found in the present investigation are consistent with the premise that the infarcted rats were in a state of moderate, compensated CHF.

Methodological considerations. 1) To avoid compromising the hemodynamic stability of the preparation, no blood samples were taken for analysis of systemic hematocrit. However, given the absence of altered blood gases and O₂ content in this model of CHF (25), we consider it unlikely that the CHF animals exhibited any alterations of systemic hematocrit. Moreover, the fluid replacement schedule during the microscopic evaluation was not different between control and CHF rats and is unlikely to have induced alterations of systemic hematocrit between groups. 2) It is possible that we did not see all capillaries within a given tissue area, particularly those without RBCs in the lumen; however, Damon and Duling (3) demonstrated that very few capillaries (<2%) fall into this category. More importantly, this latter investigation compared in vivo epifluorescence with transillumination microscopy and determined that a significant proportion (up to one-third) of capillaries containing RBCs might not be detected using transillumination microscopy (3). The present investigation did utilize transillumination microscopy, in part, to avoid the tissue damage that might result from the epifluorescence technique. Therefore, it is acknowledged that not all capillaries within the observation window may have been detected in the present investigation. Inasmuch as total capillary lineal density was greater in CHF and not less than that found in the sham spinotrapezius muscle, it is unlikely that the altered capillary lineal density reported here arose from an inability to visualize individual capillaries. Thus the results presented here are not an artifact of the microscopy technique. 3) We examined only a small area (270 × 210 µm) of tissue per screen. However, the capillary hemodynamics did not differ significantly between screens within a given muscle. Also, as mentioned above, the proportional reduction in skeletal muscle RBC flux in CHF compared with sham rats found in this investigation was consistent with the decrements in bulk flow reported by others (17, 27).

Blood flow. Although not unequivocal, the majority of the literature reports that CHF reduces blood flow 10-50% in skeletal muscle of rats (6, 27) and humans (17, 38, 42) at rest. Despite using different techniques, our measurements of tissue RBC flux reductions in CHF compared with sham muscle are in close quantitative agreement with those cited previously. In our investigation the deleterious changes in RBC hemodynamics can be attributed to the CHF condition and not a detraining effect per se, given that sham and CHF rats were housed in small, individual cages, which limited spontaneous activity (35). Simonini et al. (35) determined that CHF rats housed in this fashion exhibited activity patterns that were not different from those of sham-operated controls. Moreover, we found that the severity of these hemodynamic aberrations was increased with the severity of the myocardial infarction. Specifically, rats with infarcts <35% demonstrated a 27% reduction in RBC delivery per unit fiber width, whereas rats with infarcts >35% demonstrated a 65% decrease compared with sham control values. These decreases in RBC flux are quite similar to the decrements in skeletal muscle bulk blood flow reported previously. Indeed, Musch and Terrell (27) reported a 21 and 54% reduction in hindlimb blood flow of rats with infarcts <30 and >30%, respectively.

Arteriolar O₂ diffusion. There is evidence (16) that there may be, under certain circumstances, significant O₂ diffusion from the arterioles primarily to the surrounding tissue and capillaries. However, the measurement of RBC flux remains a valid index of O₂ delivery to the muscle as a whole, although the arteriole may assume an important role in the exchange process. With respect to the comparison between control and CHF animals, it is likely that this precapillary O₂ loss will be exacerbated in the low-blood-flow conditions present in CHF.

Hemodynamics. The percentage of capillaries supporting continuous RBC flux was significantly reduced in the CHF rats. Given the extent of these hemodynamic derangements, one remarkable feature of the CHF muscles was that capillary tube hematocrit was not statistically different from sham values. This is indicative of a similar proportion of capillary surface area being available for O₂ and substrate exchange in individual continuously RBC-perfused capillaries of sham and CHF skeletal muscle.

Theoretical basis for hemodynamic alterations. Given that the capillary does not appear to be the site of blood flow limitation (no obvious blockages or reduction in luminal diameter were observed), why do fewer capillaries remain continuously RBC-perfused at rest in the CHF condition? We speculate that there are at least three possible considerations. First, and well-documented under normal, physiological conditions, the proportion of capillaries sustaining RBC perfusion may be controlled at the arteriolar level and subject to an increased arteriolar tone due to increased levels of ANG II, endothelin, and catecholamines in conjunction with decreased availability of nitric oxide in CHF (15, 44). Second, it could simply be that the increased venous congestion and/or constriction impairs upstream capillary RBC flow by reducing the pressure differential across the capillary bed (9, 24, 44). Third, given the heterogenous RBC velocity distribution in skeletal muscle (10, 29), it is possible that a small proportion of capillaries normally sustaining the slowest RBC flow in sham animals may experience stopped flow in the CHF condition. Within these capillaries in CHF animals, RBC flow may become so slow that the interaction between the RBC and either the lumen wall or endothelial cell glycocalyx actually stops flow. It is possible that the reduction in percentage of RBC-perfused capillaries is a protective mechanism. Specifically, by reducing the proportion of RBC-perfused capillaries in CHF (potentially via one of the mechanisms described above), the muscle is able to constrain the full magnitude of the fall in RBC velocity and flux that would result when bulk blood flow is reduced up to 50% (6, 17, 27, 38, 42).

Capillary geometry. The slightly (although significantly) lower c(K,0) in the CHF spinotrapezius muscles could not be accounted for on the basis of sarcomere length differences between groups. The mechanistic basis for this effect is uncertain; however, it is relevant that other pathological conditions that cause capillary involution, such as type I diabetes, are associated with substantially straighter capillaries at a given muscle sarcomere length (13). CHF (43) and type I diabetes (13) can induce capillary involution and muscle fiber atrophy, raising the possibility that the more tortuous capillaries are lost or the remaining capillaries are under greater longitudinal "stretch" consequent to fiber structural alterations. It should be kept in mind that this effect is very modest in size and results in only a mean 3% reduction in the length of individual capillaries. This is well within the range of normal variability (at any given sarcomere length) described for healthy muscle (22).

Modeling capillary O₂ exchange. According to the theory developed by Federspiel and Popel (7), the skeletal muscle capillary surface area per capillary available for O₂ exchange will be a function of capillary tube hematocrit and capillary length. Capillary tube hematocrit was unchanged in CHF rats. However, the straighter, more highly oriented capillaries [i.e., c(K,0) = 1.10 and 1.07 in sham and CHF, respectively] found in the CHF rat spinotrapezius muscles suggest that capillary length was decreased by ~3%. Thus, within each RBC-perfused capillary, Dm_O2 should be reduced only slightly by CHF. Moreover, inasmuch as RBC velocity (and blood flow) is reduced by ~40%, capillary RBC residence time must increase and, with it, the potential for increased O₂ extraction in the spinotrapezius muscle capillaries of CHF rats. By using Gray's data (9) for capillary length within the spinotrapezius (430 µm) and correcting this on the basis of the c(K,0) measurements obtained here, mean capillary lengths of 473 and 460 µm are obtained for the sham and CHF rats, respectively. For the RBC velocity values reported here (i.e., 252 and 144 µm/s for sham and CHF, respectively), total capillary RBC residence time will be ~1.9 s for sham and 3.2 s for CHF muscles. When capillaries supporting intermittent RBC flow are included, the residence time is unchanged for sham and increased to 3.3 s for CHF rats. It is acknowledged that this value may somewhat underestimate true residence time, because the actual capillary RBC path length may be somewhat longer than the anatomic path length (32). This could explain partially the increased O₂ extraction reported in CHF patients at rest and at a given submaximal workload (30, 42), consistent with an elevated ratio of Dm_O2 to blood flow (see introduction and Ref. 31). This was demonstrated by Longhurst et al. (19) in CHF patients, where, at rest, increased O₂ extraction compensated for reduced forearm blood flow, thus maintaining resting O₂. However, during static exercise, forearm O₂ was significantly reduced in CHF compared with healthy controls. According to Fick's law (assuming a constant arterial O₂ saturation), the reduced RBC flux will necessitate an increased O₂ off-loading from each RBC to sustain a given O₂. If Dm_O2 properties are not enhanced in CHF (and there is no suggestion that they are), intracellular PO₂ must fall. One consequence of a reduced intracellular PO₂ would be impaired energetics, i.e., increased intracellular free ADP levels (41) and an enhanced stimulation of glycolysis and lactic acid production. At equivalent levels of exercise or O₂ demand, CHF patients and animals accumulate greater blood lactate concentrations and a reduced pH (21, 30, 40).

Conclusion. This investigation has demonstrated, for the first time, the effect of CHF on skeletal muscle capillary hemodynamics. CHF decreases the proportion of continuously RBC-perfused capillaries in the spinotrapezius muscle at rest. This is accompanied by a reduction in RBC velocity and flux in those vessels supporting RBC flow. Given that capillary geometry and tube hematocrit are not substantially altered, the reduced capillary RBC velocity will elevate the ratio of Dm_O2 to blood flow within the RBC-perfused capillary and this provides a likely mechanistic basis for the elevated O₂ extraction characteristic of skeletal muscle in this disease state. Moreover, the absence of changes in capillary tube hematocrit in CHF indicates that alterations in blood-tissue exchange found in this condition do not result from any capillary hematocrit-induced alteration of Dm_O2.